Converting Crypto-Mining GPU Hardware to AI Inference
Assess whether former crypto-mining GPU hardware can become useful AI capacity, including compatibility, memory, host, cooling, security and benchmark gates.
Buyer’s reading room
Private AI server guides for decisions that survive scrutiny.
Begin with the decision in front of you. Check sources, dated commercial inputs and unresolved hardware questions before relying on a recommendation.
Help and information centre
Research organised by the decision in front of you.
Follow a learning path or open the topic group that resolves the current uncertainty. Review dates identify when prices, rules and software versions need checking again.
- Method
- Evidence and claims policy
- Library state
- 12 reviewed guides
Understand and size 5 topics
Build the technical foundation before choosing a product: how old mining equipment translates to AI, where memory goes, what multi-GPU means and how serving optimisations change the result.
How Much VRAM Do You Need for a Local LLM?
Size GPU memory for a local LLM using model weights, quantisation, context, KV cache, concurrency and the difference between per-GPU and aggregate VRAM.
Tokens, Latency, Throughput and Concurrency Explained
Understand time to first token, inter-token latency, throughput and concurrency so AI server benchmark figures describe a real service.
Multi-GPU PCIe Topology, NVLink and Interconnects
Understand aggregate VRAM, PCIe paths, NUMA placement, NVLink and NVSwitch before choosing a multi-GPU AI workstation or rack server.
Quantisation, Batching, Caching and Speculative Decoding
Understand four AI inference optimisation techniques, what each changes and why benchmark conditions must remain visible.
Test and buy 4 topics
Compare runtime routes, build a reproducible acceptance test and confirm the site before accepting a hardware recommendation.
vLLM vs SGLang vs TensorRT-LLM vs llama.cpp
Compare four inference runtime routes by checkpoint support, hardware, operations, API, optimisation and reproducible benchmark evidence.
How to Benchmark and Choose an AI Server
Build an acceptance set, benchmark exact AI server configurations and choose by quality, latency, capacity, power, operations and evidence.
Private AI Server Buyer’s Guide for UK Organisations
A practical UK buyer’s guide to private AI servers: workload, model fit, VRAM, power, security, supplier evidence, costs and the cloud alternative.
AI Server Power, Cooling and Noise: UK Site Guide
Plan a UK site for a GPU or AI server: circuits, kW, electricity, cooling overhead, heat, noise, racks, UPS and pre-delivery checks.
Govern and fund 2 topics
Use these notes to frame UK data-protection, VAT and capital-allowance questions. They point to primary sources and specialist advice, not a guaranteed treatment.
UK GDPR and Private AI: Local Hosting Limits
A practical guide to UK GDPR and private AI, covering lawful basis, data routes, DPIAs, rights, retention, security, accuracy and shared responsibility.
AI Server Tax, VAT and Capital Allowances UK
Guarded UK guidance on AI server VAT, Annual Investment Allowance, full expensing, leasing and why capital gains or R&D claims should not drive a purchase.
Optional spare capacity 1 topic
Read this only after the business workload and operating costs stand on their own. Marketplace income is optional upside, never the base purchase case.
Renting Spare GPU Capacity: Options and Risks
A guarded guide to renting spare business GPU capacity through Vast.ai, Render or decentralised compute networks, including revenue, security and cost gates.
Before relying on a guide
Check dates, sources and tests before buying
Check the primary source
Regulatory, platform and technical claims can change after the stated review date.
Commercial inputs carry dates
Energy, prices and exchange rates are snapshots. Replace them before a purchase decision.
Test the exact build
A package, image or reference configuration does not prove performance. Confirm the exact build with a recorded acceptance test.
Reading should connect back to the physical system
Check how the room, hardware, management interface and application data path affect the design. Supplier reference images and technical illustrations are labelled separately.
